From conversation to insight: competitive learning and generative AI for customer experience extraction in the financial sector
Academic Article in Scopus
Purpose ¿ Customer experience (CX) is a key driver of cost efficiency, revenue growth, and competitive advantage in the banking industry. From a service-dominant logic and relationship marketing perspective, experience emerges through repeated interactions in which value is co-created over time. However, traditional tools such as surveys and Net Promoter Score (NPS) provide only partial insights into value-in-use and relationship value and often fail to capture customers' lived experiences. This study addresses this limitation by proposing a data-driven approach that directly captures customer experience from unstructured conversational data. Design/methodology/approach ¿ The study combines competitive learning and generative artificial intelligence (GenAI) to analyze 10, 062, 406 customer¿agent conversations collected monthly by a financial institution between July 2023 and July 2025. The proposed system allows customer experience specialists to query a conversational agent in natural language and receive context-rich responses, enabling exploratory customer experience analysis. Findings ¿ The results show that a conversational agent created through competitive learning and GenAI yields richer, more actionable insights into customer experiences than conventional methods. The framework demonstrates strong validity based on human evaluation and AI benchmarking against Gemini and DeepSeek (human scores >4.4; similarity >0.9 with Gemini; similarity >0.85 with DeepSeek). Compared with BERTopic and Top2Vec, it achieves higher computational efficiency, lower costs, and scalable, interpretable outputs for managerial use. Originality/value ¿ This study is among the first to apply competitive learning and GenAI to customer experience analytics in banking, offering a scalable pathway to improve service performance and competitive positioning.
